Instance failures and Auto Scaling Group behavior
The Amazon EC2 Auto Scaling Group (ASG) monitors the Snapshot Manager EC2 instance periodically. The ASG determines the status of the instance using the default status checks or via custom health checks. After the instance passes the status checks, the instance is marked as healthy. If the state of the instance changes due to an external event, for example--if a disaster causes a loss of the instance, or if you manually stop the instance, the ASG immediately marks the instance as unhealthy and schedules it for a replacement.
The ASG creates a new instance from the same Amazon Machine Instance (AMI) and uses the same configuration as that of the original instance that was configured earlier. The ASG creates a new EBS volume using the snapshot, attaches that volume to the new EC2 instance, and then brings Snapshot Manager up on that instance. You can suspend the health check process if you do not want ASG to replace the instance, for example in cases where you want to stop the instance for maintenance purposes.
For more information on how the Amazon EC2 Auto Scaling works, refer to the following Amazon AWS documentation: